Oligonucleotides or oligonucleotide analogs designed to inactivate selected single-stranded genetic sequences unique to a target pathogen were first reported in the late 1960's by Belikova, 1967, and subsequently by: Pitha, 1970; Summerton, 1978a,b, 1979a,b; Zamecnik, 1978; Jones, 1979; Karpova, 1980; Miller, 1979, 1980, 1985; Toulme, 1986; and Stirchak, 1987, 1989. Polymeric agents of this type achieve their sequence specificity by exploiting Watson/Crick base pairing between the agent and its complementary single-stranded target genetic sequence. Because such polymers only bind single-stranded target genetic sequences, they are of limited value where the genetic information one wishes to inactivate exists predominantly in the double-stranded state.
For many pathogens and pathogenic states duplex genetic sequences offer a more suitable target for blocking genetic activity. One of the earliest attempts to develop a sequence-specific duplex-directed nucleic acid binding agent was reported by Kundu, Heidelberger, and coworkers during the period 1974 to 1980 (Kundu 1974; Kundu 1975; Kundu 1980). This group reported two monomeric agents, each designed to hydrogen-bond to a specific base-pair in duplex nucleic acids. However, these agents were ineffective, probably for two reasons. First, they utilized a nonrigid ambiguous hydrogen-bonding group (an amide) which can act as either a proton donor or acceptor (in the hydrogen-bonding sense). Secondly, they provided an insufficient number of hydrogen bonds (two) for complex stability in aqueous solution. Experimental results from a variety of systems suggest that hydrogen-bonded complexes are stable in aqueous solution only if there are a substantial number (probably at least 12) of cooperative intermolecular hydrogen bonds, or if there are additional stabilizing interactions (electrostatic, hydrophobic, etc.).
Another early attempt was reported by Dattagupta and Crothers at Yale and coworkers in Germany (Kosturko 1979; Bunemann 1981). These workers employed a polymer prepared from a dye known to intercalate into duplex DNA rich in G:C base-pairs and another dye which preferentially binds to duplex DNA rich in A:T base-pairs, probably via minor-groove sites. Preparation of the polymer involved modification of the two dyes by adding acrylic moieties and then polymerization of a mixture of the modified dyes in the presence of duplex DNA of defined sequence (the template). The expectation was that the resultant polymer would show a specific affinity for duplex DNA having the same sequence as the template DNA. However, such material proved to exhibit only nominal sequence specificity. A variety of bis-intercalating agents designed to bind to specific sequences in duplex DNA have also been reported (Pelaprat, 1980), but such agents inherently give only minimal sequence specificity.
More recently, Dervan has taken a natural B-form-specific minor-groove-binding antibiotic (Distamycin) and systematically extended its structure to achieve a significant level of sequence specificity (Schultz 1982; Schultz 1983; Youngquist 1985). He has also appended to this oligomer an EDTA/Fe complex which under certain conditions acts to cleave the duplex target sequence near the agent's binding site. However, this particular approach will not lead to the high level of specificity which is needed for therapeutic applications because the inherent symmetry of the H-bonding sites in the minor groove provides too little sequence information.
Still more recently, Dervan and coworkers reported a binding agent which utilizes the informationally-richer polar major-groove sites of a target genetic duplex for sequence-specific recognition (Sluka 1987). This entailed adapting a synthetic polypeptide, comprising the DNA-sequence-recognition portion of a DNA-binding protein, for cleaving DNA at the protein's binding site on duplex DNA. The cleaving activity was achieved by linking an EDTA/Fe complex to the amino terminus of the synthetic peptide and demonstrating that this complex selectively cleaved duplex DNA at or near the parent protein's natural target sequence.
Another approach to duplex targeting has grown out of studies first reported in the late 1950's that demonstrated, via X-ray diffraction, that under high salt conditions an all-thymine or all-uracil polynucleotide can bind to specific polar major-groove sites on a Watson/Crick genetic duplex having all adenines in one strand and all thymines or uracils in the other strand (Hoogsteen binding; Hoogsteen 1959). Subsequently, it was reported that in high salt and at pH values lower than 7, an all-cytosine polynucleotide, having the cytosine moieties protonated, can bind in a similar manner to a Watson/Crick duplex having all guanines in one strand and all cytosines in the other strand.
Thereafter, is was demonstrated that under high salt and at a pH below 7, a polynucleotide containing both cytosines and thymines (or uracils) can bind to a Watson/Crick duplex having the appropriate sequence of purines in one strand and pyrimidines in the other strand (Morgan, 1968).
In the 1970's this Hoogsteen binding mechanism was exploited for affinity chromatography purification of duplex genetic fragments containing runs of purines in one strand and pyrimidines in the other strand (Flavell, 1975; Zuidema, 1978). In 1987 Dervan and coworkers exploited this Hoogsteen binding mechanism to position an all-pyrimidine polynucleotide, carrying an EDTA/Fe cleaving moiety, onto a target genetic duplex having a specific sequence of purines in one strand and pyrimidines in the other strand (Moser, 1987).
A major-groove binding mode different from the Hoogsteen mode was reported in the mid-1960's and involves binding of an all-purine polynucleotide, poly(dI), to a poly(dI)/poly(rC) duplex (Inamn 1964) and to a poly(dI)/poly(dC) duplex (Chamberlin 1965). Similarly, a mostly-purine polynucleotide has been recently used by Hogan and coworkers (Cooney, 1988) for blocking the activity of a selected natural duplex genetic sequence. These workers reported that in the presence of 6 mM Mg.sup.++ a mostly-purine polynucleotide (24 purines, 3 pyrimidines) of a specific sequence inhibits transcription of the human C-myc gene in a cell-free system.
To date, reported polynucleotides used for binding to genetic duplexes fail to satisfy one or more important criteria for effective use within living organisms. Some of these criteria are as follows:
First, the Hoogsteen-binding polynucleotides (polypyrimidines) containing cytosines require a lower-than-physiological pH in order to achieve effective binding (due to the necessity of protonating the cytosine moieties), although it has recently been demonstrated by Dervan and coworkers that the use of 5-methylcytosines in place of cytosines allows Hoogsteen binding at a pH somewhat closer to physiological (Mahler, 1989), and use of both 5-methylcytosines in place of cytosines and 5-bromouracils in place of thymines (or uracils) improves binding still further (Povsic, 1989).
Second, in the case of polypurine polynucleotides, both inosine (hypoxanthine) and adeninc moieties lack adequate sequence specificity and adequate binding affinity for effective major-groove binding in intracellular applications. The inadequate sequence specificity for inosine (Inman, 1964) and adenine (Cooney, 1988) moieties derives from the fact that inosine can bind with similar affinity to the central polar major-groove sites of both a C:I (or C:G) base-pair (i.e., NH4 of C and 06 of G or I) and an A:T or A:U base-pair (i.e., NH6 of A and 04 of T or U), and because adenine can bind with similar affinity to the central polar major-groove sites of both a T:A or U:A base-pair (i.e., 04 of T or U and NH6 of A) and a G:C base-pair (i.e., 06 of G and NH4 of C), as discussed further below.
The low binding affinity of inosine for its target base-pairs and of adenine for its target base-pairs is due to the fact that these purines can form only two less-than-optimal hydrogen-bonds to the major-groove sites of their respective target base-pairs.
Third, both polypyrimidine and polypurine polynucleotides fail to achieve effective binding to their target genetic duplexes under physiological conditions, due to the substantial electrostatic repulsion between the three closely-packed polyanionic backbones of the three-stranded complexes. Although this repulsion can be attenuated by high salt (Morgan, 1968), divalent cations (Cooney, 1988), or polyamines (Moser, 1987), nonetheless, for applications in living cells, and particularly cells within intact organisms, control of intracellular cation concentrations is generally not feasible.
Fourth, for therapeutic applications polynucleotides are less than optimal because: (i) they are rapidly sequestered by the reticuloendothelial lining of the capillaries, (ii) they do not readily cross biological membranes, and (iii) they are sensitive to degradation by nucleases in the blood and within cells.
For many in vivo applications of sequence-specific duplex-directed nucleic acid-binding agents, the principal target is DNA, which appears to exist within cells predominantly in a B or B-like conformation. In this context, polynucleotides which have been used for major-groove binding to genetic duplexes (Moser, 1987; Cooney, 1988) have a unit backbone length that is shorter than optimal for binding to duplex genetic sequences existing in a B-type conformation.